Change of Basis for Linear Transformations
You'll learn how to find the matrix representation of a linear transformation with respect to any choice of bases, not just the standard ones.
The Core Idea
When we represent a linear transformation T: V → W as a matrix, that matrix depends entirely on which basis we choose for V and which we choose for W. The "same" transformation looks different in different coordinate systems—just like GPS coordinates look different from street addresses for the same location.
Given:
- A linear transformation T: V → W
- Bases B for V and C for W
- The matrix [T]_std in standard bases
We want: [T]_{B,C} — the matrix of T with respect to bases B and C.
The Change of Basis Formula
The key insight is that we can "translate" between coordinate systems using transition matrices:
[T]_{B,C} = P_C^{-1} · [T]_std · P_B
Where:
- P_B changes coordinates from basis B to standard
- P_C^{-1} changes coordinates from standard to basis C
Think of it as a three-step journey: convert input from B-coordinates to standard, apply the transformation, then convert output from standard to C-coordinates.
Why This Matters
Different bases can dramatically simplify a transformation's matrix. A messy matrix in standard coordinates might become diagonal (or nearly so) in a cleverly chosen basis—this is exactly what diagonalization exploits! The transformation itself doesn't change; only our numerical description of it does.
Key Takeaway: The matrix of a linear transformation with respect to non-standard bases is found by "sandwiching" the standard matrix between appropriate change-of-basis matrices: [T]_{B,C} = P_C^{-1} · [T]_std · P_B.